-------------------------------------------------------------------------------
March 26, 2018


DESCRIPTION OF DATASET

This folder contains the replication files for the paper �Using Internet Search Data to Examine the Relationship between Anti-Muslim and Pro-ISIS Sentiment in U.S. Counties� by Christopher A. Bail, Friedolin Merhout, and Peng Ding, which appears in Science Advances, 2018. 

These data describe the average monthly search volumes at the county level between 2014 and 2016 for pro-ISIS and anti-Muslim search phrases on Google Search and sites partnering with Google to show ads, as well as search volumes for the phrase �weather� and �halal� for the same locations and time frames. Additionally, the data include county-level control variables from the U.S. Census, the American Community Survey, the U.S. Religion Census, the Federal Communications Commission and the Defense Manpower Data Center�s Defense Casualties Analysis System. 

All variables have been processed using statistical disclosure control methods to prevent county identification because of security concerns. Attempts to replicate the findings may thus produce slightly different estimates than those reported in the article cited above.

Please cite these data using the citation listed above.


ABSTRACT (of article)

Recent terrorist attacks by second or third-generation immigrants in the United States and Europe indicate radicalization may result from the failure of ethnic integration�or the rise of inter-group prejudice in communities where �home-grown'' extremists are raised. Yet such community-level drivers are notoriously difficult to study because public opinion surveys provide poor measures of both prejudice and radicalization. We examine the relationship between anti-Muslim and pro-ISIS internet searches in 3,099 U.S. counties between 2014 and 2016 using instrumental variable models that control for various community-level factors associated with radicalization. We find anti-Muslim searches are strongly associated with pro-ISIS searches�particularly in communities with high levels of poverty and ethnic homogeneity. Though more research is needed to verify the causal direction of this relationship, this finding suggests minority groups may be more susceptible to radicalization if they experience discrimination in settings where they are isolated and therefore highly visible�or in communities where they compete with majority group members for limited financial resources. We evaluate the validity of our measures using several other data sources and discuss the implications of our findings for the study of terrorism and inter-group relations as well as immigration and counter-terrorism policies.



-------------------------------------------------------------------------------

AUTHOR�S CONTACT INFORMATION

Christopher A. Bail
Lowey Associate Professor of Sociology and Public Policy
254 Soc./Psych Hall
Duke University
417 Chapel Dr. 
Durham, NC 27708
christopher.bail@duke.edu

--------------------------------------------------------------------------------

FILE DESCRIPTION

README.txt						This file
Bail Merhout and Peng Public Use Data			Main dataset
Bail Merhout Peng Codebook				Codebook for main dataset
Final Analysis Script Bail Merhout Peng			Main analysis script 


--------------------------------------------------------------------------------

DESCRIPTION OF VARIABLES

[1] "Population_Size�     

Resident population of county, 2010. 

[2] "Unemployment_Rate"                         

Civilian labor force unemployment rate, 2010. 

[3] "Foreign_Born_Residents_Percent"                               

Proportion of residents born abroad, 2005-2009.

[4] "Adolescent_Population_Share"           

Resident population between 10 and 19 years as a share of total resident population, 2010.                

[5] "Welfare_Spending_per_Capita"                        

Direct general expenditures of local government for public welfare per capita, FY 2002.

[6] "Male_Unmarried_Rate"                          

Proportion of males 15 years and over who are never married, separated, or divorced, 2005-2009.               
         
[7] "Poverty_Rate"                

Proportion of resident population of all ages in poverty, 2009.
              
[8] "Percent_Male"       

Proportion of resident population of all ages that is male, 2010.

[9] "Herfindal_Index_Race"             

Herfindal diversity index for census race and ethnicity measures, 2005-2009.

[10] "Non_Violent_Crime_Rate"        

Numbers of robberies, property crimes, burglaries, larcency thefts, and motor vehicle thefts known to police per capita, 2008.
      
[11] "Population_Per_Sq_Mile"         

Population per square mile, 2010.
    
[12] "Muslim_Residents_Percent"        

Muslim population (Source: ARDA Religion Census) as a share or resident population, 2010.

[13] "HS_Completion_Rate"      

Proportion of resident population 25 years and over who have at least a high school diploma, 2005-2009.

[14] "Internet_Ratio"                   

Ratio of residential fixed high-speed connections over 200 kbps in at least one direction to households, 2016.

[15] "Halal_Search_Rate"                  

Average monthly volume of searches for �halal�, 2014-2016.

      
[16] "Weather_Search_Rate"     

Average monthly volume of searches for �weather�, 2014-2016.

              
[17] "Count_of_Casualties"     

Total number of casualties from Operations Enduring Freedom, Iraqi Freedom, New Dawn, Inherent Resolve, and Freedom's Sentinel.

             
[18] "Pro_ISIS_Search_Rate"

Average monthly volume of searches for the exact phrases �How to join ISIS�, �How to join the Islamic State�, �How to support ISIS�, and �How to support the Islamic State�, 2014-2016.


[19] "Anti_Muslim_Search_Rate"       

Average monthly volume of searches for the exact phrases �Muslims are terrorists�, �Muslims are bad�, �Muslims are dangerous�, and �Muslims are evil�, 2014-2016.
